The best way to handle this type of problem is to use an XSLT processor together with an XSL stylesheet to transform the XML document into whatever output you want - in this case the contents of the XML element <ErrorTextNPA>. UNIX command line tool such as awk, sed, etc were never designed to efficiently handle XML documents.
I would note that this is not possible for the supplied XML document since it is not a well-formed or valid XML document (missing root, missing opening tags, etc.)